:J. Phys. Soc..Jpn. 67 (1998) 272-279.
4. "Coexistence of, and competition between, superconductivity and magnetism in YbPd2Sn"
Y. Aoki, H.R. Sato, T.D. Matsuda, H. Sugawara, H. Sato :J. Magn. Magn. Mater. 177-181 (1998) 559-560.
5. "Specific-Heat Anomaly around Metamagnetic Transition in UCoAI"
T.D. Matsuda, Y. Aoki, H. Sugawara, H. Sato, A.V. Andreev and V. Sechovsky : J. Phys. Soc. Jpn. 68 (1999) 3922-3926.
6. "Magnetic anisotropy, first-order-like metamagnetic transitions, and large negative magnetoresistance in single-crystal Gd2PdSi3"
S.R. Saha, H. Sugawara, T.D. Matsuda, H. Sato, R. Mallik and E.V.
maran
:Phys. Rev. B 60 (1999) 12162-12165.
7. "Low temperature properties in Ce(Col_xCux)2Ge2"
K. Maeda, H. Sugawara, T.D. Matsuda, T. Namiki, Y. Aoki, H. Sato, K. Hisatake :Phvsica B 259-261 (1999) 401-402.
8. "Magnetic arid transport properties in CeAuA13 single crystal"
H. Sugawara, S.R. Saha. T.D. Matsuda, Y. Aoki, H. Sato, .J.L. Gavilano, H.R. Ott :Physica B 259-261 (1999) 16-17.
104
9. "Metamagnetic-like anomaly in C'eCuci"
Y. Aoki, T. Namiki. T.D. Matsuda. H. Sugawara. H. tto. R. Settai. Y. Onuki B 259-961(1999) 98-99.
10. "Transport properties icross the inetamagneric transition in 1:CoAl"
T.D. Mat suda. H. Sugawara. Y. Aoki. H. Sato. A. V. Andreev. Y. Shiokawa. V.
Sechovskv. L. Havela
:Physica B 259-261(1999) 2-10-211.
11. "Nletamagnetic-Like Anomalies in f- Elect roll Systems Investigated hy Specific Heat"
H. Sato. Y. Aoki, T.D. Matsuda, H. Sugawara, A.\-. Andreev. V. Sechovskv. L.
Havela. R. Settai. Y. Onuki
:Thu. Appl. Phys. Series 11 pp.9.
12. "Fermi Surface Properties in Sr,RuOi"
Y. Yoshida. A. :\Iukai. R. Settai. N. Miyake. Y. Inada. Y. Onuki. K. Betsuvaku. H.
Harima. T.D. Matsuda. Y. Aoki and H. Sato
:J. Phys. Soc. Jpn. Vol.68 (1999) 3041-3053.
13. "Specific-Heat anomaly of metamagnetism on PrFe.IPF, and LCoAl"
T.D. Matsuda. H. Okada. H. Sugawara. Y. Aoki. H. Sato. A.V. Andreev. Y. iokawn. Sechovskv. T. Houma. E. Yamamoto, Y. Onuki
:Phvsica B 981k982 (9000) 220-222.
14. "Unusual behavior in REFe,IPL, (RE: La. Pr and Nd)"
H. Sato. Y. Abe. H. Okada. T.D. .\latsuda. N. Sugawara. Y. Aoki :Physica. B 281282 (2000) 306-307.
15. "Resistivitv minimum and anisotropy in 132PdSi3 (R, = Ce. Gd)"
S.R. Saha. H. Sugawara. T.D. Matsuda. Y. Aoki. H. Sato and E.V.
maran
:Phvsica B 281k232 (2000) 116-117.
16. "Anomalous 1()w-energy excitation in CeAtiA13"
Y. Aoki. S.R. Saha. T.D. _latsuda. H. Sugawara. H. Sato B 981K-982 (2000) 110-111.
105
17 . "Metamagnetic anomaly in single-crystalline CeFe.,(;e> and Ch „La,.Fe.,Ge...
H. Suga«ara..f. Natniki. . Yuasa. T.D. Matsuda. Y. Aoki. H. Sato, N. AIuslinikov.
S. Haile. 'r. Coto
:Pli sic a I3 2S1 7252 (2000) 69-70.
IS. "Fermi surface and superconducting properties in Sr,RuO t.,
Y. Yoshida. A. 1\Iukai. N. AIivake. A. Watanabe. R. Sertai. Onuki. T.D. AIatsuda.
Y. _Aoki. H. Sato. Y. Nlivanioto. Wada :Pli~ sica. B 251 k252 (2000) 959-960.
19. "Magnetic. thermal. and transport properties of single crystals Of antiferromagnetic I-Condo-lattice C'e>P ISi3"
S.R. Saha. H. Sugawara. T.D. Matsuda. Y. Aoki. H. Sato : Phys. Rev. 13 62 (2000) -f2.3-429.
20. "Transport properties of anisotropic itinerant-electron metamagnet UCo F
T.D. Matsuda. H. Sugawara, Y. Aoki. H. Sato. A. V. Andreev. Y. Shiokawa. V.
Secliuvskv. L. Hayela
: Pliys. Rev. B 62 (2000) 1:3852-138:55.
21. "Structural and \Iagnetic Properties of RFe1P12(R = Pr. Nd) Studied by Neutron Diffract ion"
L. Keller. P. Fisher. T. Herrinannsdorfer. A. Dduni. H. Sugawara. T.D. Matsuda.
K. Abe. Y. Aoki and H. Sato.
to be published in .1. Alloys. Comp. (2001).
22. Unusual behavior in the transport properties of R.EFe.iPl (RE:La.Ce.Pr and N(IY.
H. Sato. Y. Abe. H. Okada. T.D. AIatsuda. N. Abe. H. Sugawara. Y. Aoki
: Pliys. Rev. B 62 (2000) 13123-15130.
23. -Pressure effect on the magnetic properties in PrFe 1P i•,..
T.D. A latsnda. K. _A he. F. W atanuki. T. Namiki. S.R. Saha. H. Sugawara. A . Aoki and H. Sato
:.1. AIa ;ii. MIagti. AIater.(to be published).
106
2..1.噛'O} )Servah・1ユ ・)flleaVぎele(・rr〔 川simhehll・ ・dskluterlldh・ 、PrF・;Pl2via VanAIPhme伽(ギ
H.Sugawam.T.D.Ma.rsl!da.K.Abe.Y.Aoki,H.Sa.h}、S.Nojiri.
SρttaialldY.()1111ki
:・1̲⊃ 〜.1〜二、911̲NI〜19119二 ＼1〜1r(、r.(t̀)i)(、1 ̲)11})lis}1(}(1、.
Th('d('Flaas̲
Y.hlada.R.
107
Supplement
Journal of the Physical Socicl ill Vol. 68, No 12, December, 1999.
l;~lr+n pp 3922-392(i
Specific-Heat Anomaly around Metamagnetic Transition in UCoAI
Tatsurna D. MATsUDA*. Yuji Aoi<i, Alexander V. ANDREI2V1 **
Hitoslli SUGAvVARA, f lideyltki Srro.
and Vladiniir SEcllovSKY-' *' Department of Physics, Tokyo Metropolitan. University,
Min.arnz-Ohsawa 1-1. Hachioji-shr 192-0397 t Irist.r.tu.te of Physics, ASCR, Na Slovance 2, 18040 Prague 8, Czech
2 Department of Metal Physics
, Charles University, Ke Karlova 5, 12116 Prague 2, Czech. Republic
(Received April 14, 1999)
Republic
We have measured the specific heat and magnetocaloric effect of an UCoAI single crystal down to —0.1K in fields applied along the c: axis to investigate the thermodynamic aspects of the metamagnetic behavior. Hysteresis behavior observed in both physical quantities below --3K confirms the first order character of the metamagnetic transition . We have found that an apparent decrease of the electronic part of specific heat Ce/T by 20% at 0.25 K with increasing field across the metamagnetic transition field of —1T, evidencing the change in the density-of-states at the Fermi energy. Ce/T gradually decreases with increasing temperature in zero field. This anomalous temperature dependence can be explained by taking into account a spin fluctuation contribution in the low-field state.
KEYWORDS: UCoAI, metamagnetic transition, specific heat, magnetocaloric effect, spin fluctuation
§1. Introduction
Since the discovery of the transition from a para-magnetic state to a field induced ferropara-magnetic state in the itinerant 3d-electron systems, such as YCo2 and LuCo2,1) these systems have been intensively stud-ied as itinerant-electron metamagnets.2) Recently, in some of the heavy-fermion 4f-electron systems, such as CeRu2Si2,3-5) CeCu6,6) CeNi9Ge27) and CeFe2Ge2,8) the metamagnetic-like increase of magnetization from a paramagnetic ground state has been observed and is one of the most attractive subjects in highly correlated elec-tron systems. In spite of intensive studies, the origin of the metamagnetic-like behavior in these systems is still controversial. To answer this question, it is important to compare it with the metamagnetic anomalies found in and 5f-electron metamagnetic systems. In the 3d-systems, however, some experimental techniques includ-ing specific heat measurement are impossible, because of its high transition field Bm 70 T. In the 5f-systems, UCoAI is believed to be the most suitable compound for such purpose since the transition field is relatively low (BM 1T). Although there was a debate about the type of its zero-field ground state in the earlier stage,9) it has been confirmed that the ground state is of pa.ram-agnetic.1°)
UCoAI can be viewed as a member of UTX com-pounds, where T and X represent a transition metal and a p metal element, respectively. The magnetic behavior of these compounds has been systematically and inten-sively studied to elucidate the influence of the 5f-ligand electron interaction.1°, 11) Among them, UCoAI belongs
to a subgroup which crystallizes in the hexagonal Fe2P-type structure. It consists of two different Fe2P-types of basal-plane layers, U-Co and Co-Al, stacking along the c-axis.
Reflecting the layered structure, the metamagnetic be-havior in UCoA1 is strongly anisotropic. The magnetiza-tion (M) for the field (B) along the c-axis exhibits the metamagnetic transition (MT) around 1 T (=- B5.1). For the fields in the perpendicular plane, it increases linearly up to 39 T like a conventional weak paramagnet.101
Further studies on UCoAI are needed since the mech-anism of the metamagnetic transition has not yet been fully understood. The 5f-band splitting has been in-ferred as a possible origin based on the polarized neu-tron diffraction experiments12.13) and the band struc-ture ca.lculations.14) On the other hand, the competition between the ferromagnetic and antiferrolnagnetic inter-actions was suggested to be important in UCoAI based on the investigation on the alloy systems UT 1 _ r Co,, Al (T=Fe, Ni and Ru),15, 16)
In this paper, we report the first field-dependence measurement of the specific heat along with the magnetocaloric-effect study in UCoAI. It is shown that, to investigate field induced anomalies including the metamagnetic transition, this type of experiment is a more powerful tool than the ordinary temperature-dependence measurement of specific heat.
" E -mail: IruatsudaC'plrys.nretro-u.ac. .jp
** Joint Laboratory for Magnetic Studies . Prague.
§2. Experimental
A single crystal of UCoAI was prepared by inciting the components of 99.98% U, 99.99% Co and 99.99% Al with a nominal composition of 1.1:0.95:0.95 in an arc furnace with Helium gas atmosphere.17) The starting composi-tion was selected to avoid appearance of a small magne-tization of the order of 0.01 pB/f.u. in zero field. which is usually observed in the samples prepared from the
stoi-3922
1999) 0.4
0.3
CC
0.2
0.1
0
Spccific-Ilcal Anomaly around MIclamagnet.ic
0 1 2 3 4
B (T) 5
2K -6 -10 --tr 12
16 -20
Blc-axis 3K
6 7
Fig. 1. Field dependence of magnetization on UCoAl.
chiometric (1:1:1) composition.' In order to character-ize the quality of the present sample, we have measured the magnetization. In Fig. 1, the magnetization curves at various temperatures are shown. The sample size is about 2 x 1.2 x 2.5 mm3 and the estimated demagnetiz-ing field is l x 10-3 T at B = 5 T, which has not been subtracted in Fig. 1. At 2 K. the magnetization shows a sharp metalnagnetic transition around 1 T for the field parallel to the c-axis. The transition gradually broadens with increasing temperature. No spontaneous magneti-zation can be observed at B = 0 T in this sample within the experimental accuracy.
The specific heat was measured by the quasi-adiabatic heat-pulse method using a dilution refrigerator in mag-netic fields up to 8 T and at temperatures down to 0.1 K.
Since the measurement has been done on the same sam-ple with the same configuration as the magnetization measurement, we can directly compare the results with the magnetization data. The field dependence of the spe-cific heat is measured in the following way. The sample temperature is changed to a temperature of interest To using a sub-heater attached to an addenda, which is ther-mally linked to the mixing chamber weakly. After the temperature and the applied magnetic field have been stabilized, each heat-pulse is applied. The stabilization condition employed is that the sample temperature and its time dependence should satisfy (T — Io)/To < 10-3 and d[(T — To)/To]/dt < 10--5 sec-1, respectively. The
temperature increment associated with each heat-pulse is controlled to l%. The heat capacity of the addenda, which has been carefully subtracted from the measured total heat-capacity, amounts to 20 % (200 %) of the sam-ple heat-capacity at 0.2K (7K) in zero field.
To investigate the field dependence of entropy S(B), we have measured the magnetocaloric effect (AT/AB).
Using the data of AT/AB and C(T, B), S(B) can be obtained by integrating the following equation:
( aB)T—(T)\AB)(2.1)
Here, (AT/AB) is defined as a (ATT — AT)/IABI us-ing the temperature change ATt(A 1) resultus-ing from a slow increase (decrease) in the magnetic field by =
O NE ti
U
'Transition in I1('oAl
0.4
0.3
0.2
UCoAI
(+0.2)
0.1 - OT
0u 0 10 20 30
T2 (K2)
40 50
3923
Fig. 2. C..'jT as a function of 12 in several applied magnetic fields.
The data for B > 1 '1' have been shifted vertically by the value shown in the parentheses.
0.1
0.08
0.06 0.04
T-0.02
0 0
-E . - -^
2 4
B (T) 6
1.5
t o"
o
8 0
Fig. 3. Field dependence of the electronic specific-heat. coefficient 2 (full square) and the nuclear coefficient b (open square). Dot-ted curve is a guide to the eyes for y. The solid line represents the theoretical estimation of S.
0.05 T. The sweep rate of magnetic field is 0.125 'F/min.
This procedure eliminates the contribution from eddy--current heating of the sample. The specific heat and the magnetocaloric effect are alternately measured by chang-ing field.
§3. Results and Discussion
The temperature dependence of specific heat. C in sev-eral fields is shown in Fig. 2 as a C/T vs. T2 plot. The upturn in C/T at low temperatures is due to the nuclear Schottky contribution. The specific heat in the measur-ing temperature region can be described as C„-}-Cph+CN , where the three terms represent the electronic, phonon and nuclear contribution. respectively. CN is approxi-mately described as b/T2, which is the first term of a high temperature expansion of the nuclear Schottky contribu-tion. Below 1 K. the phonon term generally expressed as Cph = 13T3 can be neglected and C can be fitted by
3924 ratsum(1 I). \l.lrsi'f ,~ ct al. ( \'ol (F),
0
E ti U
0.1
0.05
0 0.1
0.05
0 0.1
0.05
0 0.1
0.05
0
1
0
BM
1 2
B (T)
3 4 5
Fig. 4. Field dependence of C/T. Open and full circles represent . the data for increasing and decreasing fields , respectively. The dotted curve at. T = 6 K is calculated from the magnetization data based on the thermodynamical relation .
x
0.16
0.12
0.08
0.04
0
0 2 4
T (K) 6
OT/AB C/T M
8
Fig. 5. Temperature dependence of the hysteresis JB,1 ouserved at Bm derived from C/T, AT/SB and Al data , Solid curve is a guide to the eyes.
2
1.5
~ 1
0.5
0 0
o Mvs .B
• CT vs . B - C/T vs. T
5 10
T (K)
15 20
x (T)
25
Fig. 6. B-T phase diagram. Maximum in dill/dB vs . B (open circles), anomalies in C/T vs. B (full squares) , cusp in C/T vs.
T (cross) and maximum in 111/B vs. 7' are plotted . Solid line is a guide to the eyes.
C=yT+7.,.(3.1)
Figure 3 shows the field dependence of (5, which almost agrees with the theoretical calculation taking into ac-count only the Zeeman splitting of the nuclear states;18) i.e., b = (am +aCo)B2 where 6.89 x 10-6 and 1.02 x 10-5 JK/T2mol are substituted for am and CYCo, respectively.
In the high-field state (B > Bz,l), the coefficient 3 of the T3 term, which does not depend on the field, is estimated to be about 3.6 + 0.3 x 10-:1 J/K4mol. The Debye tem-perature eD corresponding to this value of ,Q is —250 K.
It should be noted that the slope in the C/T vs. T2 plot in zero field (see Fig. 2) is quite different from that in the high-field state. This is most probably attributed to some anomaly in the electronic part of the specific heat.
This point will be discussed later in detail. The coeffi-cient of electronic specific-heat y obtained by fitting the temperature dependence of C/T is also shown in Fig. 3.
In zero field, y = 73 mJ/K2mol is in good agreement with the previous report") within the experimental ac-curacy. With increasing field, the value of -y decreases gradually. To study the field dependence of y more pre-cisely, we have performed the field dependence measure-ment of C/T at several temperatures. As shown in Fig. 4, there appears a clear anomaly at BM. At 0.25 K. C/T
decreases across B14 by —20% evidencing a change of the density-of-states at the Fermi level N(EF) associ-ated with the metamagnetic transition. With increasing temperature, the shape of the anomaly changes to a peak structure located at BM. Using a thermodynamical re-lation d2M/dT2 = d(C/T)/dB, we calculated the field dependence of C/T at 6 K based on the data of M(T , B) shown in Fig. 1. The result drawn as a dotted curve in Fig. 4 agrees well with the observed C/T . This analysis demonstrates that the peak in C/T vs. B corresponds to the change in the temperature dependence of M from d2M/dT2 > 0 in the low-field state to d2 n-I/dT2 < 0 in the high-field state.
In C/T vs. B around Bm (see Fig. 4), a hystere-sis behavior can be seen at low temperatures. The widths of hysteresis (ABH) derived from C/T data are shown in Fig. 5 along with those determined from the inagnetocaloric-effect and magnetization data . This fig-ure shows a rapid development of AB1 below 3 K . The observed hysteresis behavior demonstrates that the metamagnetic transition in UCoAI is of first order. The values of Bm determined from both the specific heat and magnetization data are plotted as a field vs. tempera-ture phase diagram in Fig. 6, along with the temperatempera-ture
1999) Specific-Hcat Anomaly around where 11I/B exhibits a maximum (Tmax). Above 2 K.
Bm increases with increasing temperature , almost lin-early with the slope of dB54/dT 0 .022 T/K up to 20 K, above which the anomaly in the magnetization gradually smears out. To estimated the c1Bm/c1T below 2 K. where magnetization can not be measured using our magne-tometer. we utilize the Clausius-Clapeyron relation
c1BM AS
111
dT0 (3.2) At 0.25 K, AS^-A7T is estimated to be —0.0029 J/Kinol using the data for B = 0 T and 2 T. Assuming a temper-ature independent AM to be 0.3 µB/f .u. at low tempera-tures, dBM /dT is estimated to be 0.0017 T/K at 0 .25 K.
These facts indicate that c1BM/dT decreases with de-creasing temperature and becomes zero at T = 0 K, since AS goes to zero and AM remains finite (see Fig . 1). At 2 K, the entropy decreases by AS----0.023 J/Kmol be-tween 0 T and 2 T, which leads to dB5.1/dT = 0.014 T/K.
To determine AS at high temperatures , we can alter-natively use the measurements of magnetocaloric effect . As shown in Fig. 7(a), AT/AB at 6K shows a peak at B11 and is reversible within the experimental accu-racy; below 1 K, AT/AB data becomes difficult to be analyzed because of the substantial irreversibility due to the first order character of the transition as shown in the inset of Fig. 7(a). The field dependence of S derived from eq. (2.1) is plotted in Fig. 7(b). These data are almost consistent with the entropy calculated by integrating Ce/T up to 6 K. At 6 K, even though it is difficult to confirm that the metamagnetic transition is still of first order, if we use the Clausius-Clapeyron relation with AS = —0.036 J/Kmol roughly estimated from S(0 T) and S(2 T) and AM = 0.30 /JB/f.u., the ob-tained dBM/c1T 0.021 T/K is consistent with dBM/dT determined by the magnetization measurements.
The characteristics of the thermal properties in UCoAl can be summarized as follows;
(1) The field dependence of both C/T and S shows a decrease across Bn1 below 0.5 K.
(2) The metamagnetic transition is of first order, as evi-denced by the hysteresis and irreversible heating effects across BM below 0.5 K.
(3) The slope of the temperature dependence of C/T changes between the low- and high-field states.
All these features are apparently different from the ther-modynamic properties of CeRu2Si2.19) At low tempera-tures, C/T in CeRu2Si2 shows a sharp peak against field at around Bm and this peak splits into double peaks at higher temperatures. The entropy also exhibits a sharp peak at BM as a function of B. Actually, the metam-agnetic anomaly is not of first order in CeRu2Si2; i.e., no hysteresis has been found across BM in the thermal properties and the magnetization.19, 20) From these facts, the metamagnetic anomalies in the two compounds must have different origins.
The change of the slope in the C/T vs. T2 plot. of Fig. 2 between the low- and high-field states indicates an anomalous temperature dependence of the electronic specific heat. The electronic contribution Ce/T defined by C/T — 6/T2 — QT2 is plotted in Fig. 8 for B = 0 T
:Vietaniagneric 'Transit ion in (1(:n.AI 1
0.5
0
-0 .5 0.2
0.15
ti
0.1
0
(a)
(b)
2K
B (T)
`6 K
:t925
0.45 0.4
0.35 0.3 0.25
0.2
Fig. 7. (a) Magnetocaloric effect AT/AB at 6K: inset shows the data at 0.25 and 0.5K. (b) Entropy S obtained by integrating Ce/T vs. T at 2K and 6 K (full circles). Solid curve at 6K rep-resents S(B) calculated from the A7/AB and C data . Dotted curve at 2K is a guide to the eyes .
0 E
N C.
0.1
0.08
0.06
0.04
0.02
0
0 2
T (K)
4 6
Fig. 8. Temperature dependence of Ce/T_ A theoretical curve fitted to eq. (3.3) is drawn by the solid curve.
and B = 5 T. Apparently, C'e/T in zero field is tempera-ture dependent and decreases gradually with increasing temperature. Similar behavior in Ce/T has been exper-imentally observed in weakly and nearly ferromagnetic systems such asUAl2, CeSn3 andYCo2.21) In these com pound s, such anomaly is attributed to the contribution of spin fluctuations and the temperature dependence can be reproduced by the following theoretical model:22 4)
Ce/~ = 0171' n t + a (T/lsp)2 ln(T/Ts}: )]. (3.3) in which a ferromagnetic spin-fluctuation contribution is taken into account. Ilere rn.*/To is the zero temperature many body mass enhancement, y0 is the electronic spe-cific heat constant determined from the band structure , isF is the characteristic spin-fluctuation temperature.
392( Tatsulna I). NI,\r,l-t)•a e1 a.l. \ al. 05.
We assume Ts': - 19 K for UCoAI since the magnetic susceptibility (x) exhibits a maximum at this tempera-ture (Trnax) and, therefore, kB 1 max is believed to give an characteristic energy scale of the spin-fluctuation. The Ce/'T vs. 7' curve shown by a solid line in Fig. 8 is the best fitting curve with the coefficients 7 = 7e(rn*/rn) = 0.073 .l/K2mol and a = 7000 = 0.12 J/K2mol. This curve qualitatively reproduces the observed temperature dependence of Ce/T for B = 0 T. Assuming the ap-plicability of eq. (3.3) even in applied fields, we have found that the magnitude of the second term in eq. (3.3) starts to decrease in the low-field state with increas-ing field as demonstrated by d(C/T)/dB > 0 at 6 K
(see Fig. 4). This may be considered as a suppression of the spin-fluctuation contribution of second term by the field. Since 7 is lower in the high-field state than in the low-field state, Ce/T finally becomes lower than that in zero field. This is a possible qualitative
expla-nation of the peak structure in C/T vs. B. As shown in Fig. 4, d(C/T)/dB > 0 in the low-field state is ther-modynamically equivalent to c12?IM1/dT2 > 0, which is related to the existence of the maximum in 1(T) at higher temperatures. Therefore, the interpretation of d(C/T)/dB > 0 indicating the suppression of the spin fluctuation by the field is consistent with the fact that the maximum in X (T) is generally considered to reflect the spin-fluctuation. As shown in Fig. 6, T,-nax decreases with increasing field and the maximum in x(T) cannot be seen any more in the high-field state, where the spin fluctu-ation seems to be frozen out. Noted that the devifluctu-ation of the electrical resistivity from the T2 dependence,25) which could be attributed to the involvement of the spin fluctuation, is observed only in the low-field state. The polarized neutron diffraction experiments indicate that the magnetic moments of the uranium 5f-electrons dom-inate over those of the cobalt 3d-electrons in the magneti-zation process. 12, 13) Therefore, the 5f magnetic moments must be playing the key role in the spin fluctuation ob-served in the specific heat measurement.
§4. Conclusion
The electronic specific heat of UCoAI at low temper-atures exhibits a decrease across the critical field B11 of the inetamagnetic transition for Bile-axis. This obvi-ously indicates that the inetamagnetic transition is in-timately connected with the change of the density-of-states at the Fermi energy. The specific heat and mag-netocaloric efect data indicate that the transition is of first order. The anomalous temperature dependence of Ce/T in the low-field state evidences the dominant spin-fluctuation contribution of the 5f-electrons.
Acknowledgements
The work was supported by a Grant-in-Aid for
Scien-tific Research from the Ministry of Education. Science.
Sports and Culture Of Japan and by the \Iinistry of Ed-ucation of the Czech Republic (grant - MF; 162).
I) T. Goto, ll. Aruga Katori, 1' Sakakihara, I1. 51 talnura, K.
Fukarnichi and K. Murata. J. Appl. Phys. 76 (1994) 6682.
and references therein.
2) II. Yamada: Physica B 149 (1988) 390.
3) P. Haen, J. Flouquet, F. Lapierre, P. Lejay and G Renienvi:
J. Low Temp. Phys. 67 (1987) 391
4) A. Amato, D. Jaccard, .I. Sierra, P. Haen, P. Lejay and J.
Flouquet: J. Low Temp. Phys. 77 (1989) 195.
5) H. P. van der Sleulen, A. de Visser, .J. J NI. Franse, I. "I ..1. NI.
Berendschot, J. A. A. .1. Perenboom, 1.1. van Kempen, A Lacerda., P. Lejay and J. Flouquet: Phys. Rev. B 44 (1991)
814.
6) A. Schroder, H. G. Schlager and 11. v. Lolieysen: .1. Nlagn.
Magn. Mater. 108 (1992) 47.
7) T. Fukuhara, K. i laezawa, H. Ohkuni, .1. Sakurai, II Sato, H. Azuma, K. Sugiyama, Y. Onuki and K. Kindo: J. Phys.
Soc. Jpn. 65 (1996) 1559.
8) H. Sugawara, Y. Aoki, H. Sato, N. Mushnikov and 'F. Goto:
J. Phys. Soc. Jpn. 68 (1999) 1094
9) V. Sechovsky, L. Havela, L. Heuzil, A. V. Andreev, G. and C. Schmit.zer: J. Less-Common Metals 121 (1986) 169.
10) V. Sechovsky, L. llavela, F. R. De Boer, J. J. NI. Franse, P. A.
Veenhuizen, J. Sebek, 3. Stehno and A. V. .Andreev: Physica 142B (1986) 283 and references therein.
11) L. Havela, A. V. Andreev, V. Sechovsky, I. N. Kozlovskaya, K. Prokes, P. Jacorsky, NI. I. Bartashevich, T. Goto and K.
Kamishima: Physica B 230-232 (1997) 98.
12) 54. Wulfe, J. 51. Fournier, A. Delapalme, B. Gillon, A. chovsky, L. Havela and A. V. Andreev: Physica B 163 (1990)
331.
13) R.. J. Papoular and A. Delapalme: Phys. Rev. Lett. 72 (1994) 1486.
14) O. Eriksson, B. Johansson and NI. S. S. Brooks: J. Phys.
Condens. Matter 1 (1989) 4005.
15) V. H. Tran, R. Troc and H. Noel: J. Alloys and Compounds
262-263 (1997) 471.
16) A. V. Andreev, L. Havela, V. Sechovsky, 51. J. Bartashevich, J. Sebek, R. V. Dremov and I. K. Kozlovskaya: Philos. Mag.
B 75 (1997) 827.
17) A. V. Andreev, M. I. Bartashevich, T. Goto, K. Kamishima, L. Havela and V. Sechovsky: Phys. Rev. B 55 (1997) 5845.
18) O. V. Lounasmaa: Phys. Rev. 133 (1964) A502.
19) Y. Aoki, 'F. D. Matsuda, H. Sugawara, H. Sato, II. Olikuni, R. Settai, Y. Onuki, E. Yamamoto, Y. Haga, A. V. Andreev,
V. Sechovsky, L. Havela, H. Ikeda and K. Miyake: J. Magn.
Magn. Mater. 177-181 (1998) 271.
20) T. Sakakibara, T. Tayama, K. Matsuhira, H. Mitamura and H. Amitsuka: Phys. Rev. B 51 (1995) 12030.
21) K. Ikeda, S. K. Dhar, NI. Yoshizawa and K. A. Gschneider, Jr.:
J. Magn. Magn. Mater. 100 (1991) 292.
22) N. F. Berk and J. R.. Schrieffer: Phys. Rev. Lett,. 17 (1966) 433.
23) S. Doniach and S. Engelsberg: Phys. Rev. Lett. 17 (1966) 750.
24) W. F. Brinkman and S. Engelsberg: Phys. Rev. 169 (1968) 417.
25) A. V. Kolomiets, L. Havela and V. Sechovsky: J. Appl. Phys.
83 (1998) 6435.
PHYSICAL REVIEW BV0l.t1Mf, 02, NUMBER 21I DECEMBER Transport properties of the anisotropic itinerant-electron metamagnet UCoAI
T. D. Matsuda, H. Sugawara, Y. Aoki, and H. Sato
Department of Physics, Tokyo Metropolitan University, Mina ni-Ol carccr /-/, Hrrrluaji shi, ./alcan
A. V. Andreev
Institute of Physics, ASCR, Na Slavanic 2, 18'121 Prague S. (':-ec•h Repuhli•
Y. Shiokawa
institute far MateriaLs Research. Tohokn University, Katahira 2-1-1. 4nba-ku• Sendai, ./a/pan
V. Sechovsky and L. Havela
Department of Electronic Structures, Charles University, Ke Karlorw 5, 12116 Prague 2, ('; ech Republic (Received 28 February 2000; revised manuscript received 11 Jul) 200(1)
Magnetoresislance (MR), Hall resistivity, and thermoelectric power in U('oAl, which exhibits a metamag-netic transition (MT) at low temperatures, have been investigated. Across MT, the MR shows a steplike decrease and increase for the transverse and longitudinal geometries, respectively. The temperature dependence of MR indicates that the magnetic scattering due to the spin fluctuations largely contributes to the electrical resistivity around 17 K. The Hall coefficient and the thermoelectric power show drastic changes associated with MT. Both in the low- and high-field states, the extraordinary Hall effect dominates in the temperature dependence of Fiall coefficient. The temperature dependence of thermoelectric power can he qualitatively explained by an extended s-c/ model considering the 5]. band structure.
2000-1
A metamagnetic transition (MT) from the paramagnetic ground state has been reported in a unique .5./ electron com-pound UCoAI, which crystallizes into the hexagonal ZrNiAI type. I ' Reflecting the anisotropic 5 f ligand hybridization in the layered structure, the metamagnetic behavior is strongly anisotropic. The magnetization (M) exhibits a sharp increase at BM 1 T in the applied field (B) along [0001], while it increases linearly with field at least up to 39 T for the field in the basal plane as observed in conventional weak paramagnets. t" Recently, we reported the field dependence of specific heat in UCoAI at low temperatures.' The elec-tronic specific-heat coefficient y, which amounts to 73 mJ/K2 mol in zero field, drops by 20% across Bs. . "I'his observation indicates a change of the density of states at the Fermi level. As a possible origin of MT, a field-induced ferromagnetism due to the 5f band splitting has been in-ferred based on the polarized neutron-diffraction experiments4-5 and the band-structure calculations.6'7 In spite of these investigations, the origin of a MT is not clearly understood yet.
In this paper we report the results of measurements of magnetoresistance (MR), Hall resistivity (pH), and thermo-electric power (TEP) performed on a single crystal of meta-magnet UCoAI. A single ingot has been grown by the Czo-chralski pulling method in an argon atmosphere using a tetra-arc furnace. The starting materials were 3N5(99.95% pure) U, 3N+ Co, ON Al. From the ingot, several bar-shaped samples (typically 4 x 1.3 X 0.5 mm') have been cut out for the transport measurements. The temperature dependence of electrical resistivity for n[0001] is basically the same as was
reported in Ref. 8. Judging from the residual resistivity ratio (RRR (17.3), the quality of the present sample is better than the previous one (the RRR is --4.5).8
0163-1829/2000/62(21)/13852(4)515.00 PRB 62
The electrical resistivity (p) and the Hall resistivity were measured by the conventional dc four-probe method using KEITHLEY 182 nanovoltmeters. TEP was measured by the ordinary differential method using chromel and AuFe-normal silver thermocouples. The magnetic field was applied using a superconducting magnet with fields up to 5 T and an iron-core electromagnet up to 1.5 T. Using a quantum design superconducting quantum interference device (SQUID) mag-netometer, the magnetization has been measured on the same samples under the same geometry for direct comparison with the transport measurements.
Figure 1 shows the MR for BO 0001] in (a) the longitu-dinal (Ap11) and (b) the transverse ( _ ) geometries at se-
1.5 •
,(a)Ap=p(B)-p(0T)
1 -2 .i)K
0.5- eols
a 0
B//r//10(H)tl
-0 .5 --- 5
(b)
02.0 K
-5
• a
a ID B/Iltxx)1J 1//II1201It)K -15
•
.(a) Ap=p(B)-p(1)T)
2.0K
-4.2 K //I1110 011I()K
(b)
2.0 K
117C
B 10(X).1 1 // 17201
. ,
mai
4.2 K
-10 K B IT]
FIG. 1. MR in (a) the longitudinal ( ) and (h) the transverse (Apj ) geometries irrr B [0001], both for increasing and decreasing magnetic field; there exists a small hysteresis across a MT within the experimental accuracy.
13 852©22000 The American Physical Society